Conventional techniques by which a mobile unit autonomously detects its position include GPS (Global Positioning System), optical beacon high-precision positioning, a lane marker and the like. GPS is unsuitable for high-precision positioning in a mobile unit due to the principle of positioning performed by receiving a radiowave emitted from an artificial satellite. Moreover, optical beacon high-precision positioning cannot perform positioning at a site where an optical beacon is not installed, and for a lane marker, its infrastructure equipment cost becomes very high.
Therefore, a method using an image shot by an image capturing device included in a mobile unit is known as a method of detecting a position of the mobile unit. A method of detecting a target of a known size from an image and of calculating a distance to the target based on the size of the target in the image, and a method of detecting a plurality of targets installed at known sites from an image and of calculating a position based on a relationship among positions of the plurality of targets are known.
For example, the following technique is known. A target such as a road sign, a traffic signal or the like positioned ahead of an automobile is shot with an image sensor when the automobile is running on a road, and the image is input to a processing device. The processing device processes the shot image with an image processing unit, and classifies the type of the target. The classified target is compared with shape data stored in a knowledge base, and its tilt is corrected. Thereafter, the image processing unit calculates the size of the shot target in the image as an angle (rad), and outputs the calculated angle to a distance calculating unit. The distance calculating unit calculates a distance to the target by using the angle (rad) that indicates the size of the shot target, and data that indicates the size of the target, which is stored in the knowledge base.
Additionally, an image of a landmark installed, for example, along a route of a mobile unit is captured with an image capturing unit, and position data of three landmarks prestored in a storing unit are read if at least three or more landmarks are recognized. Moreover, a technique of calculating a relative angle among landmarks viewed from a mobile unit based on an image, and of calculating the current position of the mobile unit based on position data and the angle data is known.
Furthermore, the following technique is known. For example, in a processing unit, an outline area of a road sign is extracted after image data of the road sign captured by a moving video camera is converted into HSV color system using a hue, saturation and a value as criteria. A rectangular outline is identified by using values in directions horizontal and vertical to the extracted outline area. Then, an absolute position of the road sign is determined based on positions of a plurality of outlines identified based on a plurality of pieces of image data, of the road sign captured in time series and an absolute position of the video camera at a time point when each of the plurality of pieces of image data is captured. Note that HSV stands for Hue, Saturation and Value.    [Patent Document 1] Japanese Laid-open Patent Publication No. 08-219775    [Patent Document 2] Japanese Laid-open Patent Publication No. 2001-142532    [Patent Document 3] Japanese Laid-open Patent Publication No. 2003-85535    [Patent Document 4] Japanese Laid-open Patent Publication No. 2006-53775